Activity Number:
|
672
- Methods for Infectious Disease Epidemiology
|
Type:
|
Contributed
|
Date/Time:
|
Thursday, August 2, 2018 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Section on Statistics in Epidemiology
|
Abstract #327112
|
|
Title:
|
Model Choice and Future Prediction Accuracy in Time Series for Disease Incidence
|
Author(s):
|
Reagan Spindler and Yew-Meng Koh*
|
Companies:
|
Hope College
|
Keywords:
|
Dengue Fever;
Time Series;
Bayesian;
Forecast Accuracy;
Neural Network
|
Abstract:
|
One of the goals of time series models for disease incidence data is to predict accurately future disease counts. Many such models exist, some of which utilize information from covariates. The utility of an appropriately lagged covariate is highlighted. We introduce a Bayesian neural network time series model for predicting dengue fever incidence in Singapore, which utilizes Singaporean precipitation data as a covariate. A comparison is made between this neural network model and a time series model which does not use any covariate information. A method for choosing between the models which optimizes future prediction accuracy is suggested as well.
|
Authors who are presenting talks have a * after their name.